Pith. sign in

REVIEW 4 major objections 6 minor 60 references

SVasP: Self-Versatility Adversarial Style Perturbation for Cross-Domain Few-Shot Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that averaging localized crop style gradients into the global adversarial style perturbation stabilizes optimization and flattens the loss landscape, improving cross-domain few-shot transfer beyond the prior StyleAdv…

desk verdict Plausible incremental extension of StyleAdv with consistent gains that are likely inflated by tuning hyperparameters on the eight target test sets. read the letter →

arxiv 2412.09073 v1 pith:K2Z2Z7FX submitted 2024-12-12 cs.CV cs.LG

Wenqian Li , Pengfei Fang , Hui Xue This is my paper · ORCID
classification cs.CVcs.LG
keywords cross-domainfew-shotlearningsinglesourcedomainstyleperturbationadversarialtraininggradientstabilityflatminimaself-versatilitygeneralization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Single-source cross-domain few-shot learning asks a model trained on one labeled domain to recognize new classes in visually different domains, and style-based methods perturb image style statistics to bridge the gap. The paper claims that the previous best such method, StyleAdv, suffers from unstable gradient directions and converges to narrow, source-specific minima, and that both problems can be attacked together by using local crop gradients. Its proposed SVasP randomly crops each image, computes style gradients for each crop and for the whole image, averages the crop gradients, and adds a small weighted amount to the global style gradient before generating the adversarial style. The authors argue this 'self-versatility' stabilizes the update direction and guides training to flatter minima, which transfers better to unseen domains. They report average 5-way 5-shot accuracies of 65.09% on ResNet-10 and 72.84% on ViT-small, versus StyleAdv's 63.77% and 72.44%.

What carries the argument

The load-bearing object is the SV Gradient Ensemble, named 'self-versatility' because all gradient information comes from within one training image. Style is defined as the channel-wise mean and standard deviation of feature maps, and the attack is performed on these statistics. For each of the first three backbone blocks, SVasP collects classification-loss gradients with respect to the style statistics of k random crops and of the whole image, averages and normalizes the crop gradients, and mixes them into the global gradient with weight xi = 0.1. That ensembled gradient determines the adversarial style through signed updates with randomized coefficients kappa1 and kappa2, and the style is transplanted with AdaIN. A second mechanism, the Discrepancy & Consistency Optimization, pushes adversarial features away from the source domain through a domain discriminator while keeping global, crop, and adversarial features semantically consistent.

What would settle it

Train SVasP with the published recipe and hold out several target datasets that were never used in any hyperparameter selection; if its advantage over StyleAdv disappears or reverses on those held-out domains, the claim that the stabilized crop perturbation transfers to unseen domains is not supported. A sharper test is to replace the averaged real crop gradients in the ensemble with random unit vectors of the same scale, and if accuracy stays the same, then diversification alone rather than the crop-gradient direction would explain the gain.

Watch

Extended reading notes

Core claim

The central discovery is that the gradient of a style-based adversarial attack, in isolation, is an unstable and incomplete guide, and that local views of the same image can supply the missing directional information. For each image, SVasP computes the mean and standard-deviation style gradients for k random crops and for the global image; the crop gradients are normalized and averaged, then added to the normalized global gradient with a small decay factor xi. The resulting ensemble gradient is used to build adversarial styles with randomized step sizes and Gaussian noise, applied through AdaIN feature normalization. The paper asserts that this stabilized perturbation makes training converge to a flatter minimum of the loss landscape, and that the gain is measurable as higher few-shot accuracy across eight target domains while keeping the source-domain classifier and FSL classifier working together.

Load-bearing premise

The whole gain rests on the assumption that gradients from two local crops, mixed into the global style gradient at weight 0.1, capture domain-agnostic structure rather than noise, and that the hyperparameter settings fitted on the eight target datasets will hold for unseen target domains.

Editorial extensions

If this is right

  • On the BSCD-FSL benchmark with ResNet-10, SVasP reports a 5-way 5-shot average of 65.09% versus StyleAdv's 63.77%, with the largest gains on ISIC, EuroSAT, and CropDisease.
  • On ViT-small pretrained on ImageNet-1K, SVasP reports 72.84% versus StyleAdv's 72.44% for 5-way 5-shot, and 59.36% versus 58.57% for the fine-tuned 1-shot setting.
  • The stabilized gradient direction is claimed to persist through training, with higher gradient cosine similarity across epochs than StyleAdv, and the loss landscape near the converged solution is flatter.
  • Ablation results attribute the gain to all three components: the SV gradient ensemble alone moves the average from 62.07% to 62.61%, adding the domain discrepancy loss reaches 63.69%, adding the consistency loss reaches 64.05%, and all together reach 65.09%.
  • Under fine-tuning, the method also improves the ResNet-10 5-shot average to 68.30% compared with StyleAdv's 67.28%.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The stabilizing effect of crop gradients is a general hypothesis, not one limited to few-shot learning; if true, similar local-gradient ensembling should improve other adversarial augmentation and domain generalization methods.
  • Because the hyperparameters k = 2, xi = 0.1, and the crop scale range were selected using the eight target datasets, the strongest evidence for the paper's claim will come from applying the fixed recipe to new target domains that were never used in tuning.
  • The claim that k = 2 is optimal suggests a bias-variance tradeoff in gradient estimation: too few crops leave the stabilizer noisy, while too many dilute the global gradient and encourage source overfitting; this could be tested by measuring gradient cosine similarity as a function of k.
  • A sharper test of the mechanism would replace the averaged real crop gradients with random unit vectors of the same scale; if accuracy is unchanged, then diversification alone, rather than the specific crop-gradient direction, would explain the gain.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes SVasP, a method for single-source cross-domain few-shot learning that augments global style perturbation with gradients computed from randomly cropped local image regions. The method combines a Self-Versatility gradient ensemble (averaging crop style gradients into the global style gradient), an adversarial style perturbation step, and a Discrepancy & Consistency Optimization (DCO) objective with domain-discrimination, crop-consistency, and global-adversarial consistency losses. The authors claim that this stabilizes gradient updates and drives the model to flat minima, improving transfer to unseen target domains. Experiments are reported on the BSCD-FSL and mini-CUB benchmarks with ResNet-10 and ViT-small backbones, showing average accuracy gains over StyleAdv and other baselines, along with ablations, hyperparameter studies, loss-landscape visualizations, and Grad-CAM maps.

Significance. If the empirical results are trustworthy, the central idea—using localized crop style gradients as stabilizers for global style perturbation—is a plausible and potentially useful contribution to CD-FSL. The paper includes extensive experiments, per-dataset tables, ablations, and a public code link, which are strengths. However, the significance is materially weakened by three issues: hyperparameters are selected on the eight target test datasets used for final evaluation, the flat-minima claim rests only on qualitative visualizations, and the ablation attributes most of the improvement to auxiliary losses rather than to the proposed gradient-ensemble mechanism. The reported margins over the strongest baseline are also small, especially for ViT-small, and no significance tests are given. These concerns limit the strength of the paper's central claims in its current form.

major comments (4)
  1. [Implementation Details; Qualitative Evaluation (Figures 3–5, Tables 5–8)] The key hyperparameters—number of crops k=2, decay factor ξ=0.1, consistency weight λ=0.2, crop scale s=(0.2,0.4), and the choice κ1≠κ2—are selected by maximizing accuracy on the same eight target datasets that are later used for final evaluation. Figures 3–5 and Tables 5–8 report per-dataset accuracy on ChestX, ISIC, EuroSAT, CropDisease, CUB, Cars, Places, and Plantae for each hyperparameter setting and pick the best average. Because these datasets are the test sets for the main comparisons in Tables 1–2, the reported gains (e.g., +1.32 on ResNet-10 5-shot, +0.40 on ViT-small 5-shot) are optimistically biased and do not establish generalization to genuinely unseen domains. The paper should adopt a validation protocol that uses only source-domain data (or a separate validation set) for hyperparameter selection, then fix those hyperparameters and report the resulting test-set numbers.
  2. [Qualitative Evaluation; Figures 1, 6, 10] The central claim that SVasP 'escapes from poor sharp minima' and 'obtains a flattened minima' is supported only by a schematic (Figure 1(b)), qualitative loss-landscape plots for BSCD-FSL and mini-CUB (Figures 6 and 10), and a gradient-cosine-similarity curve (Figure 1(a)). The loss-landscape plots compare only 'without SV gradient' and SVasP, not the actual baseline StyleAdv, and no quantitative flatness measure (e.g., dominant Hessian eigenvalue, sharpness metric, or PAC-Bayes bound) is provided. Since the flat-minima mechanism is a headline contribution of the abstract and introduction, quantitative evidence is needed to make this claim load-bearing.
  3. [Table 3 (Ablation study)] The ablation does not support the attribution of the gains to the proposed self-versatility gradient ensemble. Table 3 shows that adding the SV module alone to the baseline improves average accuracy from 62.07% to 62.61% (+0.54%), whereas the DCO losses Ldom and Lcon add the remaining 2.48% (from 62.61% to 65.09%). The text states that the SV Gradient Ensemble module 'effectively utilizes the source domain style gradients,' but the numbers suggest the auxiliary objectives, not the stabilized global style perturbation, dominate the improvement. To substantiate the mechanism claim, the paper should include a full-model-minus-SV condition (i.e., baseline + Ldom + Lcon without SV) and, if possible, an experiment that isolates the gradient-stabilization effect while keeping the DCO losses fixed.
  4. [Tables 1 and 2 (Quantitative results)] The claim that SVasP 'significantly outperforms' existing state-of-the-art methods is not supported by statistical evidence. The ViT-small 5-shot improvement over StyleAdv is only 72.84% vs. 72.44% (+0.40%), and the ResNet-10 5-shot average gain of +1.32% is reported without a confidence interval on the average. The per-dataset confidence intervals often overlap between SVasP and StyleAdv (e.g., ChestX 5-shot 26.87±0.38 vs. 26.07±0.37). The authors should report confidence intervals for the average accuracy and, where appropriate, paired significance tests across the eight datasets, so that the reader can assess whether the margins are meaningful.
minor comments (6)
  1. [Eq. (20) and surrounding text] The text says 'We use Kullback-Leibler divergence loss KL(·) to maximize global-adversarial consistency,' but Ladv is added to the total loss and is minimized, which makes the adversarial and global predictions closer. The wording 'maximize' is inconsistent with the optimization direction; please rephrase.
  2. [Figure 1(a) and Introduction] The figure caption describes the gradient cosine similarity 'between epochs,' while the introduction says 'we measure the gradient cosine similarity between the forward and backward gradients.' Please clarify which quantity is plotted and define it precisely in the text.
  3. [Implementation Details] The 'style change probability' is set to 0.2 but is not included in any ablation or sensitivity study. Please state how this value was chosen and whether it is a fixed hyperparameter or tuned.
  4. [Table 2] The PMF baseline has entries only for ChestX, ISIC, EuroSAT, and CropDisease in the 5-shot block, while other cells are dash-marked. Please specify whether those datasets were not evaluated for PMF or whether the numbers were omitted.
  5. [Methodology; Equation (17)–(19)] The relation classifier f_re and the FSL predictions p_fsl_i are not described in sufficient detail. The main text mentions 'GNN' and 'ProtoNet' as classifiers, but the architecture and training of f_re used for the crop consistency loss are not specified; please add details.
  6. [Conclusion / Related Work] The phrase 'this is the inaugural study exploring the impact of localized style gradients on model generalization' is a strong historical claim. Please temper it to a more precise statement about contribution relative to the cited prior work.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method's losses, gradient updates, and ablations are self-contained; test-set hyperparameter selection is an evaluation-leakage concern, not a circular derivation.

full rationale

Walking the claimed derivation chain, SVasP's components are defined independently of the target results: style statistics are computed by Eqs. (1)-(2), crop and global gradients are computed from standard cross-entropy losses in Eqs. (5)-(7), averaged and normalized in Eqs. (8)-(11), and used as sign-based adversarial perturbations in Eqs. (12)-(15). The DCO losses (18)-(20) are standard domain-adversarial and consistency objectives. None of these equations define the claimed outcome (improved target transfer) in terms of that outcome; nor does the paper invoke a self-citation as the load-bearing justification for its core mechanism. The concern that crop-gradient stabilization is inferred only from full-pipeline success is not a circular step: it is an empirical attribution question, and the ablation in Table 3 partially addresses it. The hyperparameters (k, xi, lambda, s, kappa1/kappa2) are selected using the eight target test datasets in Figures 3-5 and Tables 5-8, which weakens the claim that the targets are truly unseen for model selection and can inflate the reported margins; however, this is test-set selection bias rather than a derivation that reduces to its own inputs by construction. No self-definitional, self-citation-load-bearing, or uniqueness-imported step was found.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The method relies on several empirical assumptions about styles and minima, and its main hyperparameters were selected using target test data. No new physical or conceptual entities are introduced beyond the self-versatility framing.

free parameters (7)
  • decay factor ξ = 0.1
    Blends crop gradients into the global gradient (Eq. 10-11); swept over {0,0.01,0.1,1} and best value chosen by target-domain accuracy in Table 5.
  • number of crops k = 2
    Number of random crops used for gradient ensemble; optimal value chosen from {0,...,5} based on target-domain accuracy in Figure 3 and Table 8.
  • consistency weight λ = 0.2
    Balances global-crop consistency and crop FSL loss in Eq. (19); chosen by target-domain sweep in Table 6.
  • attack step sizes κ1,κ2 = [0.008, 0.08, 0.8]
    Step sizes for style perturbation in Eqs. (14)-(15); randomly selected from this set, not swept over other ranges.
  • Gaussian noise scale ε = 16/255
    Initial perturbation noise in Eqs. (12)-(13); set to a standard adversarial noise magnitude.
  • crop scale s=(sl,sh) = 0.2, 0.4
    Bounds for random cropping; chosen from a grid based on target-domain accuracy in Figure 5.
  • style change probability = 0.2
    Probability of applying style perturbation during training; fixed without a dedicated ablation.
assumptions (4)
  • domain assumption Style statistics (mean and variance) capture domain identity.
    The method assumes that perturbing these statistics simulates unseen domains, following StyleAdv and MixStyle.
  • domain assumption Adversarial sign-gradient steps in style space improve transferability.
    The gradient ascent on style parameters is borrowed from adversarial training; no proof is given for cross-domain transfer.
  • domain assumption Flatter minima generalize better.
    The paper claims flat minima from stabilized gradients, invoking the sharp-flat minima connection without theoretical justification.
  • domain assumption Random crops preserve semantic labels.
    Crop inputs are assigned the same labels as the global image in the consistency losses (Eq. 19).

how reviews work

0 comments
Cite this review

Pith. "Pith review of SVasP: Self-Versatility Adversarial Style Perturbation for Cross-Domain Few-Shot Learning." pith.science (2026). https://pith.science/paper/K2Z2Z7FX

@misc{pith2026241209073,
  author       = {Pith},
  title        = {Pith review of: SVasP: Self-Versatility Adversarial Style Perturbation for Cross-Domain Few-Shot Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/K2Z2Z7FX}},
  note         = {Machine review of arXiv:2412.09073}
}
read the original abstract

Cross-Domain Few-Shot Learning (CD-FSL) aims to transfer knowledge from seen source domains to unseen target domains, which is crucial for evaluating the generalization and robustness of models. Recent studies focus on utilizing visual styles to bridge the domain gap between different domains. However, the serious dilemma of gradient instability and local optimization problem occurs in those style-based CD-FSL methods. This paper addresses these issues and proposes a novel crop-global style perturbation method, called \underline{\textbf{S}}elf-\underline{\textbf{V}}ersatility \underline{\textbf{A}}dversarial \underline{\textbf{S}}tyle \underline{\textbf{P}}erturbation (\textbf{SVasP}), which enhances the gradient stability and escapes from poor sharp minima jointly. Specifically, SVasP simulates more diverse potential target domain adversarial styles via diversifying input patterns and aggregating localized crop style gradients, to serve as global style perturbation stabilizers within one image, a concept we refer to as self-versatility. Then a novel objective function is proposed to maximize visual discrepancy while maintaining semantic consistency between global, crop, and adversarial features. Having the stabilized global style perturbation in the training phase, one can obtain a flattened minima in the loss landscape, boosting the transferability of the model to the target domains. Extensive experiments on multiple benchmark datasets demonstrate that our method significantly outperforms existing state-of-the-art methods. Our codes are available at https://github.com/liwenqianSEU/SVasP.

Figures

Figures reproduced from arXiv: 2412.09073 by the authors.

Figure 1
Figure 1. SVasP stabilizes the gradients and escapes from [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our proposed methods SVasP. “RB” is an abbreviation for ResNet Block. Random cropping the benign [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Performances on different numbers of crops [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Performances on (a) different ξ, λ and (b) whether use same κ1, κ2. The average accuracy (%) is reported. Accuracy ChestX 26 27 25 0.1 0.3 0.5 0.7 ISIC 51 49 0.1 0.3 0.5 0.7 EuroSAT 0.1 0.3 0.5 0.7 87 85 89 Accuracy Places 74 76 78 0.1 0.3 0.5 0.7 𝑠௛ ൌ 0.4 𝑠௛ ൌ 0.5 𝑠௛ …
Figure 5
Figure 5. Figure 5: Performances on different scale parameters [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 8
Figure 8. Figure 8: Comparison of style generation methods for dif [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Comparison of the vanilla global perturbation and our self-versatility perturbation methods. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Loss landscape visualization results of the model [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Grad-CAM visualization results of the StyleAdv [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 48 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Caron, M.; Touvron, H.; Misra, I.; J \'e gou, H.; Mairal, J.; Bojanowski, P.; and Joulin, A. 2021. Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9650--9660

    work page 2021

  4. [4]

    Chen, C.; Tang, L.; Liu, F.; Zhao, G.; Huang, Y.; and Yu, Y. 2022. Mix and reason: Reasoning over semantic topology with data mixing for domain generalization. In Proceedings of Advances in Neural Information Processing Systems, 33302--33315

    work page 2022

  5. [5]

    F.; and Huang, J.-B

    Chen, W.-Y.; Liu, Y.-C.; Kira, Z.; Wang, Y.-C. F.; and Huang, J.-B. 2018 a . A Closer Look at Few-shot Classification. In Proceedings of International Conference on Learning Representations

    work page 2018

  6. [6]

    Chen, Z.; Badrinarayanan, V.; Lee, C.-Y.; and Rabinovich, A. 2018 b . Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks. In Proceedings of International Conference on Machine Learning, 794--803

    work page 2018

  7. [7]

    Das, D.; Yun, S.; and Porikli, F. 2021. ConfeSS: A framework for single source cross-domain few-shot learning. In Proceedings of the International Conference on Learning Representations, 1--12

    work page 2021

  8. [8]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. In Proceedings of the International Conference on Learning Representations, 1--12

    work page 2020

Show all 60 references
  1. [9]

    Du, Z.; Li, J.; Su, H.; Zhu, L.; and Lu, K. 2021. Cross-domain gradient discrepancy minimization for unsupervised domain adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 3937--3946

    2021

  2. [10]

    Feng, F.; Wang, J.; and Geng, X. 2024. Transferring Core Knowledge via Learngenes. arXiv preprint arXiv:2401.08139

    2024 arXiv

  3. [11]

    Feng, F.; Wang, J.; Zhang, C.; Li, W.; Yang, X.; and Geng, X. 2023. Genes in Intelligent Agents. arXiv preprint arXiv:2306.10225

    2023 arXiv

  4. [12]

    Feng, F.; Xie, Y.; Wang, J.; and Geng, X. 2024 a . Redefining <Creative> in Dictionary: Towards a Enhanced Semantic Understanding of Creative Generation. arXiv preprint arXiv:2410.24160

    2024 arXiv

  5. [13]

    Feng, F.; Xie, Y.; Wang, J.; and Geng, X. 2024 b . Wave: Weight template for adaptive initialization of variable-sized models. arXiv preprint arXiv:2406.17503

    2024 arXiv

  6. [14]

    Fu, Y.; Fu, Y.; and Jiang, Y.-G. 2021. Meta-fdmixup: Cross-domain few-shot learning guided by labeled target data. In Proceedings of the ACM International Conference on Multimedia, 5326--5334

    2021

  7. [15]

    Fu, Y.; Xie, Y.; Fu, Y.; Chen, J.; and Jiang, Y.-G. 2022. Me-d2n: Multi-expert domain decompositional network for cross-domain few-shot learning. In Proceedings of the ACM International Conference on Multimedia, 6609--6617

    2022

  8. [16]

    Fu, Y.; Xie, Y.; Fu, Y.; and Jiang, Y.-G. 2023. Styleadv: Meta style adversarial training for cross-domain few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 24575--24584

    2023

  9. [17]

    Garcia, V.; and Bruna, J. 2018. Few-shot learning with graph neural networks. In Proceedings of the International Conference on Learning Representations, 1--12

    2018

  10. [18]

    C.; Karlinsky, L.; Codella, J

    Guo, Y.; Codella, N. C.; Karlinsky, L.; Codella, J. V.; Smith, J. R.; Saenko, K.; Rosing, T.; and Feris, R. 2020. A broader study of cross-domain few-shot learning. In Proceedings of the European Conference on Computer Vision, 124--141

    2020

  11. [19]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 770--778

    2016

  12. [20]

    Helber, P.; Bischke, B.; Dengel, A.; and Borth, D. 2019. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7): 2217--2226

    2019

  13. [21]

    Hu, Y.; and Ma, A. J. 2022. Adversarial feature augmentation for cross-domain few-shot classification. In Proceedings of the European Conference on Computer Vision, 20--37

    2022

  14. [22]

    R.; Panda, R.; Karlinsky, L.; Feris, R.; and Radke, R

    Islam, A.; Chen, C.-F. R.; Panda, R.; Karlinsky, L.; Feris, R.; and Radke, R. J. 2021. Dynamic distillation network for cross-domain few-shot recognition with unlabeled data. In Proceedings of Advances in Neural Information Processing Systems, 3584--3595

    2021

  15. [23]

    Kim, T.; and Han, B. 2024. Randomized adversarial style perturbations for domain generalization. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2317--2325

    2024

  16. [24]

    Krause, J.; Stark, M.; Deng, J.; and Fei-Fei, L. 2013. 3d object representations for fine-grained categorization. In Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, 554--561

    2013

  17. [25]

    Laenen, S.; and Bertinetto, L. 2021. On episodes, prototypical networks, and few-shot learning. In Proceedings of Advances in Neural Information Processing Systems, 24581--24592

    2021

  18. [26]

    Li, H.; Xu, Z.; Taylor, G.; Studer, C.; and Goldstein, T. 2018. Visualizing the loss landscape of neural nets. In Proceedings of Advances in Neural Information Processing Systems, 1--11

    2018

  19. [27]

    Li, L.; Xiao, J.; Chen, G.; Shao, J.; Zhuang, Y.; and Chen, L. 2024. Zero-shot visual relation detection via composite visual cues from large language models. In Proceedings of Advances in Neural Information Processing Systems

    2024

  20. [28]

    Li, P.; Gong, S.; Wang, C.; and Fu, Y. 2022 a . Ranking distance calibration for cross-domain few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9099--9108

    2022

  21. [29]

    Li, X.; Dai, Y.; Ge, Y.; Liu, J.; Shan, Y.; and DUAN, L. 2022 b . Uncertainty Modeling for Out-of-Distribution Generalization. In International Conference on Learning Representations, 1--13

    2022

  22. [30]

    Liang, H.; Zhang, Q.; Dai, P.; and Lu, J. 2021. Boosting the generalization capability in cross-domain few-shot learning via noise-enhanced supervised autoencoder. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 9424--9434

    2021

  23. [31]

    Liu, B.; Zhao, Z.; Li, Z.; Jiang, J.; Guo, Y.; and Ye, J. 2020. Feature transformation ensemble model with batch spectral regularization for cross-domain few-shot classification. arXiv preprint arXiv:2005.08463

    2020 arXiv

  24. [32]

    P.; Hughes, D

    Mohanty, S. P.; Hughes, D. P.; and Salath \'e , M. 2016. Using deep learning for image-based plant disease detection. Frontiers in Plant Science, 7: 215232

    2016

  25. [33]

    P.; and Hariharan, B

    Phoo, C. P.; and Hariharan, B. 2021. Self-training For Few-shot Transfer Across Extreme Task Differences. In International Conference on Learning Representations, 1--13

    2021

  26. [34]

    Qi, L.; Dong, P.; Xiong, T.; Xue, H.; and Geng, X. 2024. DoubleAUG: Single-domain Generalized Object Detector in Urban via Color Perturbation and Dual-style Memory. ACM Transactions on Multimedia Computing, Communications and Applications, 20(5): 1--20

    2024

  27. [35]

    Rame, A.; Dancette, C.; and Cord, M. 2022. Fishr: Invariant gradient variances for out-of-distribution generalization. In Proceedings of the International Conference on Machine Learning, 18347--18377

    2022

  28. [36]

    Ravi, S.; and Larochelle, H. 2017. Optimization as a model for few-shot learning. In Proceedings of the International Conference on Learning Representations, 1--11

    2017

  29. [37]

    R.; Cogswell, M.; Das, A.; Vedantam, R.; Parikh, D.; and Batra, D

    Selvaraju, R. R.; Cogswell, M.; Das, A.; Vedantam, R.; Parikh, D.; and Batra, D. 2017. Grad-cam: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 618--626

    2017

  30. [38]

    Snell, J.; Swersky, K.; and Zemel, R. 2017. Prototypical networks for few-shot learning. In Proceedings of Advances in Neural Information Processing Systems, volume 30

    2017

  31. [39]

    Sun, J.; Lapuschkin, S.; Samek, W.; Zhao, Y.; Cheung, N.-M.; and Binder, A. 2021. Explanation-guided training for cross-domain few-shot classification. In Proceedings of the International Conference on Pattern Recognition, 7609--7616

    2021

  32. [40]

    H.; and Hospedales, T

    Sung, F.; Yang, Y.; Zhang, L.; Xiang, T.; Torr, P. H.; and Hospedales, T. M. 2018. Learning to compare: Relation network for few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1199--1208

    2018

  33. [41]

    Triantafillou, E.; Zhu, T.; Dumoulin, V.; Lamblin, P.; Evci, U.; Xu, K.; Goroshin, R.; Gelada, C.; Swersky, K.; Manzagol, P.-A.; et al. 2020. Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examples. In International Conference on Learning Representations, 1--13

    2020

  34. [42]

    Tschandl, P.; Rosendahl, C.; and Kittler, H. 2018. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Scientific Data, 5(1): 1--9

    2018

  35. [43]

    Tseng, H.-Y.; Lee, H.-Y.; Huang, J.-B.; and Yang, M.-H. 2020. Cross-domain few-shot classification via learned feature-wise transformation. In Proceedings of the International Conference on Learning Representations, 1--14

    2020

  36. [44]

    Van Horn, G.; Mac Aodha, O.; Song, Y.; Cui, Y.; Sun, C.; Shepard, A.; Adam, H.; Perona, P.; and Belongie, S. 2018. The inaturalist species classification and detection dataset. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 8769--8778

    2018

  37. [45]

    Wah, C.; Branson, S.; Welinder, P.; Perona, P.; and Belongie, S. 2011. The caltech-ucsd birds-200-2011 dataset

    2011

  38. [46]

    Wang, H.; and Deng, Z.-H. 2021. Cross-domain few-shot classification via adversarial task augmentation. In Proceedings of the International Joint Conference on Artificial Intelligence, 1075--1081

    2021

  39. [47]

    Wang, W.; Duan, L.; Wang, Y.; En, Q.; Fan, J.; and Zhang, Z. 2022. Remember the difference: Cross-domain few-shot semantic segmentation via meta-memory transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7065--7074

    2022

  40. [48]

    Wang, X.; and He, K. 2021. Enhancing the transferability of adversarial attacks through variance tuning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1924--1933

    2021

  41. [49]

    Wang, X.; Peng, Y.; Lu, L.; Lu, Z.; Bagheri, M.; and Summers, R. M. 2017. Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. In Proceedings of the IEEE/CVF Conference on Computer Visio...

    2017

  42. [50]

    Xie, Y.; Feng, F.; Wang, J.; Geng, X.; and Rui, Y. 2024. Kind: Knowledge integration and diversion in diffusion models. arXiv preprint arXiv:2408.07337

    2024 arXiv

  43. [51]

    Yu, T.; Kumar, S.; Gupta, A.; Levine, S.; Hausman, K.; and Finn, C. 2020. Gradient surgery for multi-task learning. In Proceedings of Advances in Neural Information Processing Systems, 5824--5836

    2020

  44. [52]

    Zhang, J.; Song, J.; Gao, L.; and Shen, H. 2022. Free-lunch for cross-domain few-shot learning: Style-aware episodic training with robust contrastive learning. In Proceedings of the ACM International Conference on Multimedia, 2586--2594

    2022

  45. [53]

    Zhang, X.; Xu, R.; Yu, H.; Zou, H.; and Cui, P. 2023. Gradient norm aware minimization seeks first-order flatness and improves generalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 20247--20257

    2023

  46. [54]

    ZHENG, H.; Wang, R.; Liu, J.; and Kanezaki, A. 2023. Cross-Level Distillation and Feature Denoising for Cross-Domain Few-Shot Classification. In Proceedings of the International Conference on Learning Representations, 1--12

    2023

  47. [55]

    H.; and Sebe, N

    Zhong, Z.; Zhao, Y.; Lee, G. H.; and Sebe, N. 2022. Adversarial style augmentation for domain generalized urban-scene segmentation. Advances in Neural Information Processing Systems, 35: 338--350

    2022

  48. [56]

    Zhou, B.; Lapedriza, A.; Khosla, A.; Oliva, A.; and Torralba, A. 2017. Places: A 10 million image database for scene recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(6): 1452--1464

    2017

  49. [57]

    Zhou, F.; Wang, P.; Zhang, L.; Wei, W.; and Zhang, Y. 2023. Revisiting prototypical network for cross domain few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 20061--20070

    2023

  50. [58]

    Zhou, K.; Yang, Y.; Qiao, Y.; and Xiang, T. 2020. Domain Generalization with MixStyle. In Proceedings of the International Conference on Learning Representations, 1--12

    2020

  51. [59]

    Zhuo, L.; Fu, Y.; Chen, J.; Cao, Y.; and Jiang, Y.-G. 2022. Tgdm: Target guided dynamic mixup for cross-domain few-shot learning. In Proceedings of the ACM International Conference on Multimedia, 6368--6376

    2022

  52. [60]

    Zou, Y.; Liu, Y.; Hu, Y.; Li, Y.; and Li, R. 2024. Flatten Long-Range Loss Landscapes for Cross-Domain Few-Shot Learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 23575--23584

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.